Editorial

Abstract

Introduction

The Issues

Dynamic mapping tools

cdv

Alternatives to choropleth maps

Conclusions

Future directions

Acknowledgements

References

Case Studies Index

7. Future Directions & Technical Considerations

Collaboration and dissemination were important aspects of Project Argus. Working relationships were developed with a variety of organisations to share ideas and data. Links with the KINDS Project at Manchester Computing (URL 5) led to the notion of combining interactive graphics for data exploration with the distribution of census data from a National data server. The potential for an Internet based visual interface to the UK Census, or Visualization Gateway, is currently under investigated through Project JANUS at Manchester Computing (URL 6). The project has three main objectives towards which work is progressing:

1. Transferring the key deliverables of the ARGUS and KINDS projects into an integrated service and hosting them on MIDAS.
2. Educating the community by promoting, using and supporting these products.
3. Developing software that will enable users to perform exploratory spatial data analysis of census data effectively forming a visualization gateway to the UK Census of Population over the Internet.

The ARGUS resources have been successfully housed on the MIDAS machine. The cdv software has been used in workshops demonstrating exploratory visualization of census data to academics, researchers, statisticians, and civil servants in the UK and Europe. It has also been used with locally appropriate data in undergraduate and postgraduate geography programs at the Universities of Leeds, Keele & Leicester. A series of generalised polygons and population cartogram bases is being developed for enumerated zones, specifically for visualization with cdv or similar products.

Technological considerations relating to the final stage are currently being assessed. An attractive option would be a Java application combining the 2D and 3D census mapping. Java provides much of the functionality of Tcl/Tk, and is quicker, but at the expense of the high-level flexibility and hands-on nature of the scripting language. It is, however, a language for Internet applications and couples neatly with the VRML script node meaning that interactive querying and linking can be applied to the three dimensional representations in a similar way to the 2D interaction provided by cdv (Moore et al., 1997). The release of the Java 3D API in the spring of 1998, with its 3D graphics, will extend possibilities further. Whilst cdv was a demonstrator, for which a hands on, flexible, scripted approach was ideal, the visualization gateway will be an application, for which the speed, security and robustness of a compiled application are more appropriate. These considerations are influenced by developments initiated by Sun Microsystems to integrate Java and Tcl applications. Jacl and Tcl-Blend enable Tcl commands to be written in Java and Java programs to call Tcl procedures. These capabilities are extremely recent, and at the time of writing their utility for adding processing power to spatial Tcl applications remain untested. They might provide the means for combining the kinds of 3D graphics required to visualize a population surface with the kinds of dynamic graphics provided by cdv in two dimensions. These developments, along with the resolution of some of the Internet security issues that both languages currently address by omitting functionality, would enable an integrated visualization gateway to be developed for census data using a combination of the two languages. Both authors are currently engaged in a project that uses Tcl/Tk to add a map interface to panoramic photographs and VRML/Java to display three-dimensional data sets. Combination of the approaches is under investigation.

8. Acknowledgement

The figures used to illustrate this Case Study are based on Census data and zone boundary data for the 1981 UK Census of Population for the County of Leicestershire. These data are Crown copyright. Permission to display and distribute these, given to the JISC/NTI Project ARGUS based at Leicester University and Birkbeck College London, from the Office of National Statistics and the ED-LINE Consortium is gratefully acknowledged.